Rice is one of the most strategic crops in the world, Africa and particularly in Ethiopia1. In Ethiopia, the second most populous nation in Sub-Sahara Africa, rice is one of the target commodity that have received due emphasis in promotion of agricultural production and is considered the "Millennium crop". Ethiopia has considerably vast suitable ecologies for rice production but unsuitable for production of other food crops2.
The expansion of rice production in Ethiopia is tackled by many different constraints, with salinity as the major one. According to Hawando3, 36% of the land in Ethiopia is affected by salinity. About 40% of the irrigable area in middle awash is out of production due to salinity4. The most efficient strategy which can increase the productivity of the saline land is the use of tolerant varieties5,6.
Even though rice is known to be highly susceptible to salinity there are genotypes that can survive and produce grain at high salinity level7. The effect of salinity depends on the type of crop, environmental condition, biotic stress and growth stage5. Salinity causes stress through osmotic and specific ion effect8 that results in low germination, stunted seedling growth, high sterility, dried leaf and low grain yield of rice9-11.
Even though Ethiopia is an ideal country for production of rice, the amount of arable land under rice cultivation is very small as compared to the potential and a large quantity of rice is still imported annually12. Since, salinity is expanding in irrigated agriculture of Ethiopia4 access to new varieties tolerant to soil stress conditions, especially to salinity is an urgent task to be undertaken. Therefore, the present study was proposed to identify high yielding, salt tolerant rice genotypes.
MATERIALS AND METHODS
Description of the study site: The study was conducted at Werer Agricultural Research Center (WARC) in lath house from August-February, 2015. Werer is located 9°27 N and 40°15 E in North Eastern part of Ethiopia about 280 km from Addis Ababa. The soil in the region is predominantly vertisol with the porosity and bulk density (0-25 cm depth) of 49.06% and 1.35 g cm2, respectively3.
Treatments and designs: Factorial experiment consists of 15 rice genotypes including one susceptible check and one tolerant check at 4 levels of salinity (0, 4, 8 and 12 dS m1). The salt concentrations were selected based on the experiment by Dawit13. The genotypes were developed by IRRI (International Tice Research Institute) for salt tolerance and distributed for testing for the year 2012 growing season.
Soil collection and preparation: Soil was collected from WARC research field. The soil was air dried for 10 days followed by gently dispersing and mixing thoroughly then sieving through a 2 mm sieve. The pH value, CEC and EC of the soil were 8.1, 18 mEq/100 g soil and 1.01 dS m1. The texture of the soil was silty clay containing 40% clay, 49% silt and 11% sand.
Pot experiment management: The experiment was conducted by sowing rice seeds in plastic pots of 22 cm top diameter, 15 cm bottom diameter and 23 cm depth filled with 5 kg soil in 3:1 ratio of the collected soil and sand, respectively. Each pot was sown with 10 seeds and lined with double layer of cotton to restrict seepage of the solution. Then the pots were kept in the lath house under sunlight. Then after, the soil in the pots were moisturized with water and commercial NaCl with 12.8, 25.6 and 38.4 g were added to obtain 3 salinity levels (4, 8 and 12 dS m1). Salt solution seepage from each pot was collected every 24 h and returned to the pot to avoid loss of salts. The soil was fertilized with 50 N and 25 P mg kg1 of soil according to IRRI recommendation14. Watering and other agronomic managements were done according to the requirement of the plant.
Data collection: The data collected was based on both individual plant and pot basis.
Data collected on pot basis: The weight of 100 selected grains from each pot was taken and conversion was made to 1000 grain weight; for pots having less number of grains conversion was made according to the number of seed available. The grain weight of each pot (from central spike and tiller spike) was taken after the moisture content was adjusted to 14% moisture content using moisture content meter (Agratronix, Japan).
Data recorded on individual plant basis: Data of the following traits were recorded from five plants in the pot. Plant height was measured from the ground level to the tip of central panicle. Panicle length was measured from the bottom of the panicle to the top most fertile spikelet on the main axis. Total number of spikelets from each selected central panicles were counted and number of unfilled spikelets were counted from the same panicles. Then spikelet fertility percentage was calculated by dividing the mean number of filled spikelets to the total number of spikelets multiplied by hundred. Total number of tillers were counted from randomly selected plants and then from the same plants effective number of tillers were counted by identifying tillers that have grains.
To compare the response of different genotypes for salinity stress, various indices have been used, such as tolerance index (Tol)15, Mean Productivity Index (MPI)15, geometric mean productivity index (GMP)16, Stress Tolerance Index (STI)16, Stress Susceptibility Index (SSI)17, Yield Index (YI)18 at stress condition and Yield Stability Index (YSI)19:
where, Ypi is the mean yield of each cultivar in the control treatment, Ysi is the mean yield of each cultivar in 12 dS m1 saline treatment, Yp is the mean yield of all cultivars in control treatment and Ys is the mean yield of all cultivars in 12 dS m1 saline treatment.
Data analysis: The data was subjected to analysis of variance (ANOVA) of the two factors Completely Randomized Design (CRD) using GLM procedure of Statistical Analysis System (SAS) version 9.0. Based on the grain yield data univariate stability parameters such as Wrickes ecovalence and Shuklas stability using SAS20. Multivariate analytical tool GGE biplot were also employed to asses similarity and dissimilarity among four salinity levels and interaction patterns between genotypes and salinity levels.
RESULTS AND DISCUSSION
The names of some genotypes had been shortened for simplification. Computations of regression analysis between the mean of the parameters and the salinity levels and data between the correlations of the stress indices were not showed.
Analysis of variance: The analysis of variance showed that there is very highly significant difference between the salinity levels, the genotypes and their interaction for the parameters taken.
Yield and yield components
Plant height: The analysis of variance showed the plant height of different rice genotypes were significantly affected by the different salinity levels. This result showed the increase in salinity concentration affected the plant height negatively, that caused 21-61% decrease at 12 dS m1 even 100% for some susceptible genotypes (Fig. 1a) in agreement with the study of Hakim et al.9, who stated that the reduction in plant height is proportional with the increment in salinity concentration but tolerant genotypes can retain their height even at higher salinity concentration by outperforming the sensitive ones. It also complies with the experimental observation of Dawit13. Generally, it was observed that salinity caused the decrease in plant height in all the genotypes that could be the cumulative effect of salinity in delaying emergency, the decrease in shoot and root biomass. The reduction in plant height in the increased salinity level could be due to lower water potential and reduction in leaf water content which results stomatal closure that limits carbon dioxide assimilation and reduced photosynthetic rate12. A disturbance in mineral supply (excess/deficiency) which induce changes in the concentration of specific ions that affect the growth10 might be the other reason for reduction of plant height.
Number of tillers: Grain yield of rice is highly dependent on the number of panicle producing tillers. All the genotypes in this experiment were highly affected by the increase in salinity on total and effective number of tillers. In case of genotype IR 29 (susceptible check), IR 59418, IR 72593, IR 73055 and NERICA 4 showed 100% loss at 12 dS m1 (Fig. 1b). Minimum reduction of effective tillers were observed in IR 71810, IR 71901, IR 71991 and IR 70023 across the four salinity levels (Fig. 1b). As the salinity concentration becomes higher and higher the reduction in number of effective tillers per plant was also higher21. The same was true for this experiment, which at higher salinity concentration the number of effective tillers decreased significantly over the control. This result is supported by the study of Putech and Modal21.
||(a) Plant height of the 15 rice genotype in four salinity levels and (b) Number of tillers of the 15 rice genotype in four salinity levels, *Susceptible, **Tolerant
Panicle length: This experiment revealed that panicle length of the different rice genotypes was greatly affected by salinity. The highest and the lowest panicle length was registered in IR 71889 (Control) and the four genotypes (12 dS m1), respectively. Salinity has negative effect on the length of panicle, thus there is a decrease in the length of the panicle with an increase in the concentration of salinity. The same thing was observed in this experiment but the decreasing rate was varied with genotypes. There was very little reduction of panicle length in the tolerant genotypes like IR 70023 and IR 71810 which lost 23% at the higher salinity level and much higher in the susceptible ones like NERICA 4 and IR 59418 (Fig. 2a). This finding is supported by the experiment of Hakim et al.10, Rad et al.22 and Mahmood et al.23 where the major cause of reduction in panicle length was the reduction in seedling survival rates and stunted growth caused by salinity.
Spikelet fertility percentage: Spikelet fertility is important component of grain yield. Generally, all the genotypes showed decrease in spikelet fertility with increasing of salinity concentration. It was severe in IR 59418 (52-100%), IR 72593 (58-100%), IR 73055 (62-100%) and NERICA 4 (51-100). The IR 70023, IR 71991, IR 71901 and IR 71810 were relatively tolerant (8-30% reduction) for change in spikelet fertility with the increase in concentration of salinity (Fig. 2b). The percentage fertility of spikelet was negatively affected by the increase in salinity concentration. The same was true in all most all the genotypes in this experiment where the increase in salinity level severely affected the spikelet fertility percentage. This result is in line with the results of Dawit13 where the reduction in spikelet fertility was reported at higher salinity concentrations.
Thousand kernel weight: The IR 73055, IR 72593, IR 59418 and NERICA 4 had the lowest thousand kernel weight ranging from 7.17-10.10 g (Table 1). These genotypes could not resist the salinity concentration and did not produce any grain at 12 dS m1 but IR 71901, IR 70023, IR 71810, AT 401 and IR 71991 were tolerant with 22-37% reduction over the control.
||(a) Panicle length of the 15 rice genotype in four salinity levels and (b) Spikelet fertility of the 15 rice genotype in four salinity levels, *Susceptible, **Tolerant
The highest amount of reduction at 12 dS m1 was registered in IR 59418, IR 72593, IR 73055 and NERICA 4 where their reduction in thousand kernel weight at 8 dS m1 was more than 50%. At 8 and 12 dS m1 the lowest reduction was registered in IR 70023, IR 71810 and IR 71901 with the range of 12-18 and 22-29%, respectively (Table 1). Generally, in every addition of 1 U (dS m1) salinity between 0-12 dS m1 decreased the grain weight by 0.74 g. The influence of salinity on thousand grain weight was high in this experiment and it is in line with the result of Puteh and Mondal21 and Aref24 where grain weight decreased significantly with increasing salinity concentration.
Grain yield: Grain yield is the final sum of all components at different stage thus, the salinity effect on each traits may directly or indirectly affect the final grain yield. The increment of salinity form the control to 12 dS m1 significantly reduced the grain yield of all the genotypes. In all the salinity levels IR 59418, IR 72593, IR 73055 and NERICA 4 showed the susceptibility of salinity reducing form 59-100% yield loss between the control and the highest salinity levels (Table 1). But at 12 dS m1 IR 70023, IR 71810, IR 71901 and IR 71991 can be recommended as tolerant genotype which only showed 37-46% yield loss compared to the control. The IR 55179, IR 71889, IR 71902 and IR 72048 were moderately tolerant which showed relatively good performance at lower salinity levels (4 and 8 dS m1) by losing 12-26% at 4 dS m1 and 41-47% for IR 55179 and IR 71889 at 8 dS m1 (Table 1).
The analysis of variance also showed that there was very highly significant (p<0.001) difference between the four salinity levels of the rice genotypes on their grain yield. The highest mean grain yield (10.31 g) was registered at the control salinity treatment (0 dS m1) and the lowest mean grain yield (2.94 g) was registered at 12 dS m1. The increase in salt concentration at each salinity level caused a significant reduction of grain yield, the increment from 0 dS m1 (control) to 4, 8 and 12 dS m1 caused a loss of 31, 56 and 71% in grain yield, respectively.
In this experiment salinity caused a decrease in grain yield especially when the concentration becomes higher.
|Table 1:||Thousand kernel weight and grain yield of rice genotypes on different salinity levels
|*Susceptible check, **Tolerant|
|Table 2:||Wrickes ecovalence and Shuklas stability variance for grain yield per pot
|*Susceptible check, ** Tolerant, #Total sum of square due to heterogeneity among variance|
This agrees with the Hakim et al.10 and Aref24 that observed all the tested rice varieties were inversely influenced by salinity. The study by Grattan et al.25 showed the reduction of grain yield by 12% for every unit of increase in ECe above 3 dS m1. Hakim et al.9 explained that the reduction in grain yield was due to the manifestation of the cumulative reduction of the yield components. Although, salinity caused yield loss. There were genotypes which lost less than 40% of their grain yield of their control. These genotypes are tolerant and moderately tolerant to salinity according to the study of Puteh and Mondal21.
Stability analysis: Wrickes26 defined ecovalence as the contribution of each genotype to genotype by environment interaction. Genotype with lower Wrickes ecovalence (Wi) had smaller deviation from the environmental mean indicting the stability of the genotype. Higher Wi indicates the a higher contribution of a genotype for the genotype by environment interaction which indicates the instability in the performance of the genotype across the environments.
According to Wrickes26 IR 73055, IR 66946 (tolerant check), IR 72048 and IR 71991 had the highest contribution to genotype×salinity interaction due to their higher ecovalence value (Table 2). The IR 72593, IR 59418, IR 70023, AT 401, IR 71902 had relatively very small ecovalence which indicate the stability of these genotypes. But IR 72593 and IR 59418 had lower mean yield.
According to Shukla27 a genotype is called stable if the stability of variance (σ2i) is equal to the environmental variance (σ2e) which means that σ2i = 0. A relatively large value will indicate thee grater instability of the genotype.
According to Shukla27 IR 73055 and IR 66946 (tolerant check) showed the greatest instability (Table 2). The IR 72593, IR 59418, IR 70023 and IR 71889 were relatively stable genotypes. But also IR 72593 and IR 59418 had lower mean yield.
Multivariate analysis: The GGE biplot (Fig. 3) grouped the four salinity environments in two sectors, environments in the same sector considered as a single mega environment so the lower salinity levels 0 and 4 dS m1 grouped under one mega environment and the higher salinity levels 8 and 12 dS m1 grouped in another mega environment but 8 and 12 dS m1 are more closer. The winning genotype at 0 dS m1 (control) was IR 72048 whereas at 4 dS m1 it was IR 55179.
|Table 3:||Mean values of stress tolerance indices for rice genotypes at 12 dS m1
TOL: Tolerance index, MP: Mean productivity index, GMP: Geometric mean, productivity index, STI: Stress tolerance index, SSI: Stress susceptibility index, YSI: Yield susceptibility index, YI: Yield index, *Susceptible check, **Tolerant
GGE (Genotype main effect plus genotype by environment interaction) biplot analysis of grain yield
, 1: AT 401, 2: IR 29, 3: IR 55179, 4: IR 59418, 5: IR 66946, 6: 70023, 7: IR 71810, 8: IR 71889, 9: IR 71901, 10: IR 71902, 11: IR 71991, 12: IR 72048, 13: IR 72593, 14: IR 73055 and 15 NERICA 4
At 8 and 12 dS m1 the wining genotypes were IR 71810, IR 71901 and IR 70023. The IR 71889 was found to be unresponsive genotype. The low yielding genotypes in all salinity levels were IR 73055, IR 72593, NERICA 4, IR 29 and 4 (Fig. 3).
Generally, the at lower salinity level IR 72048, IR 55179, IR 71901, AT 401 performed good and at the higher salinity level IR 71810, IR 70023, IR 71901, IR 71991 and IR 66946 (tolerant) performed good. Therefore, these genotypes could be considered tolerant at each mega environments.
Stress tolerance indices: Genotypes having higher STI show higher grain yield stability across different environments15. Therefore, IR 70023 and IR 71810 with STI of 0.68 and 0.69, respectively were the most stable genotypes across the salinity levels. However, IR 59418, IR 72593, IR 73055 and NERICA 4 with STI value of 0 were the most sensitive ones (Table 3).
According to Khan and Kabir28 lower SSI value (SSI<1) for a given genotype indicates the higher stability of the genotype in stress and non-stress environments. The IR 71991, IR 71901, IR 71810 and IR 70023 showed lower SSI value of 0.52, 0.54, 0.56 and 0.65, respectively (Table 3). Therefore, these genotypes were the most stable genotypes among tested genotypes in stress environments. Beside this, Nouri et al.29 reported that genotypes that show lower TOL values and high MP values are more tolerant to stress. So based on the lowest TOL and the highest MP values, IR 71991, IR 71901, IR 71810 and IR 70023 were found to be tolerant.
The GMP used to determine the degree of susceptibility under both stress and non-stressed conditions; genotype with higher GMP considered tolerant and high yielding16. The GMP value of IR 59418, IR 72593, IR 73055 and NERICA 4 were 0 which indicted the susceptibility of these genotype in stress environment (12 dS m1). The IR 70023 and IR 71810 had higher GMP value that shows tolerance of these genotypes for stress environment. The YSI and YI used to discriminate tolerant genotypes in stress conditions and high value shows the tolerance of the genotype28. Therefore, IR 71810, IR 70023, IR 71901 and IR 71991 were found tolerant in stress condition.
All of the indices had strong, significant and positive correlation with grain yield in exception of SSI and TOL which had negative but strong correlation. This indicates all indices are useful in discriminating genotypes that are both tolerant and stable in both stressed and non-stressed condition. The YI, YSI, YSI and SSI had more strong correlation even perfect correlation in YI with grain yield which shows this indices are more powerful. The correlation of MPI, GMP and STI was found strong and significantly positive.
The results of correlation of this study is supported by Dawit13 who reported that MP, GMP and STI are highly and positively correlated with the grain yield in stress and non-stress environments. Also significant and positive correlation was found among MPI, GMP and STI by Khan and Kabir28 on the study of heat tolerance in bread wheat. Therefore, these indices are a good indicators in identifying high yielding genotypes in saline and non-saline environments.
In conclusion, IR 70023 and IR 71810 were found tolerant. As a recommendation, the 15 rice genotypes should further be evaluated in field experiments to confirm their salinity tolerance in the real environment. Moreover, studying the tolerance mechanism of these genotypes could simplify the selection and the variability creation method. Also, there should be a breeding program that improves the salt tolerant genotypes.
Financial support of EAAP and SARD-SC rice projects is greatly acknowledged for my MSc. study. I also thank Mr. Mulat Zerihun for his support in laboratory analysis.
||Since salinity is becoming a major problem in crop growing areas of Ethiopia, there should a way to solve this problem. Along many measures to control salinity using resistant genotypes is often the best and cheapest solution. So, this experiment able to identify some tolerant genotypes with comparable salinity levels to the ground problem
||Variety shortage especially rice varieties is also one of the problem in the irrigated regions of the country so this experiment can add up on high yielding genotypes for the irrigated areas
||It also create a good chance for other breeders in variety improvement programs by identifying promising lines for variety development
||Since, rice is good source of carbohydrate and essential proteins, it can has a great impact health in related issues especially for the poor community in need